The review addressed a clear question supported by appropriate inclusion criteria. Relevant sources were searched, but the search was limited and relevant studies could have been missed. Diagnostic and statistical filters were used during the search, which could have resulted in further studies being missed. Each stage of the review process was conducted in duplicate, which reduced the risk of error and bias.
Study quality was assessed using appropriate criteria, but some relevant criteria (such as the handling of uninterpretable results) were not assessed. There were limitations with the methods of synthesis. Firstly, the model used to derive pooled estimates did not maintain the intra-study link between sensitivity and specificity, which could result in the overestimation of accuracy; more robust models were available that could produce summary estimates whilst maintaining the link between related measures. Secondly, the meta-analyses were conducted using a random-effects model with only three studies per analysis; this was an insufficient number of studies to inform the distribution of accuracy across the studies within the model.
Despite the limitations of the review and available evidence, and the uncertainty regarding the reliability of the results for the individual tests being evaluated, the authors' conclusion about the relative accuracy of transvaginal sonography with components of physical examination is likely to be reliable.